Comparison of data mining clustering algorithms

摘要

Data mining is an area of computer and information science with large perspective of knowledge discovery from large database or dataset. Various types of disciplines are available under data mining and clustering or the unsupervised learning in particular. Clustering is a division of data into similar groups; each similar group is called a cluster. Object in a cluster are similar or close to each other. Clustering algorithms can be implemented via number of different approaches. We conducted the comparison on WEKA (The Waikato Environment for Knowledge Analysis) that is open source. This paper shows that study and comparison between different clustering algorithms-partitioning method, hierarchical method and density based method. Here we have used parameter cluster instance, iterations, sum of squared errors, time taken, etc. for prediction of forest fire.